The background of this proposal rests on three observations: First, that neurofibrillary tangles occur an anatomically related set of regions, leading to the possibility that tangle formation is linked to anatomical connectivity. Second is the discovery that a HMW form of hyperphosphorylated, soluble tau can be taken up by neurons, lead to tau aggregation, and propagate trans-synaptically. This form of tau can be isolated from Alzheimer brain and transgenic mice overexpressing mutant tau. Third is the observation from human neuropathology that tangles rarely spread from medial temporal lobe to the cortex without the presence of Abeta in the cortex, consistent with parallel observations in transgenic mice. We have recently found that the presence of Abeta also leads to several potentially related phenotypes: higher levels of tau propagation, higher levels of the unique, seed competent HMW tau, and increased neuroinflammation. These observations support the idea that there is synergy between Abeta and tau in the cortex, perhaps mediated by glia, and the core of this application is to understand this interaction. Aim 1 uses biochemical measures and a biosensor cells to ask how Abeta changes tau, in mice and in humans, and if microglia play a role. Aim 2 builds a nontransgenic model of tau uptake and aggregation, and evaluates changes in tau neurotoxicity due to age and Abeta. Aim 3 uses pharmacological agents to block soluble Abeta species (BACE inhibitor), clear fibrillar Abeta (anti-Abeta immunotherapy) and doxycycline transgene suppression to reduce tau. Already we find evidence of interesting interactions: Prolonged tau transgene suppression leads to clearance of the HMW species in Tg4510 mice, but the same HMW tau species remains remarkably stable in the APP/PS1xTg4510 cross. These data pave the way for combination therapy, because both Abeta and tau therapeutics may be necessary to target these long lived HMW tau species. Our model systems also allow us to test hypotheses that have direct clinical impact: we will test the concept that Abeta initiates tau misfolding, which can propagate without Abeta participation ? i.e. the hypothesis that there is an early ?amyloid dependent? and a later ?amyloid independent? phase of the disease. Indeed, at 3 and 6 months, APP/PS1 x Tg4510 have markedly increased HMW tau activity compared to Tg4510s, but 12-month old mice no longer show a pattern of enhanced HMW tau. If, as we propose, the spread of tangles starts out dependent on Abeta, but becomes increasingly less so as more misfolded tau accumulates, it follows that anti-Abeta therapeutics would be effective only during the ?amyloid dependent? phase, and help explain the multiple failures of clinical trials of anti-amyloid agents given later in the disease. Mapping the timing of these phases would have crucial import in informing the next generation of trials.